They are not alternatives. One finds, the other keeps — and the loop between them is the account.
People treat these as a choice. They are not: they are two halves of one machine, and running either alone gives up most of the value.
| Exact match | Search Match (Discovery) | |
|---|---|---|
| What it bids on | Terms you listed, matched as typed | Whatever Apple thinks fits your app |
| Who chooses | You | Apple |
| What it is for | Keeping what already works, at a price you set | Finding terms nobody thought of |
| Share of spend | The majority | A minority — it is paying to learn |
Search Match spends a little money discovering what real people actually type. Most of it is worthless and some of it is gold — phrases you would never have written down, because you think about your app in your own words and users do not.
When one of those searches converts, it graduates: you add it as an exact-match keyword in the campaign where you set the bid. From then on you control what it costs.
That is the whole account, and it never stops running. Discovery keeps finding, Exact keeps what was found.
The moment a term graduates into Exact, it must be added as a negative keyword in Discovery. Otherwise both campaigns bid on the same search — your own money on both sides of the auction — and you pay more for traffic you were already getting.
This is the single most common structural mistake in a self-managed account, and it is silent: nothing breaks, the cost just creeps. Apple allows a large number of negative keywords per ad group (five thousand), so there is no reason to be sparing.
Published guidance usually puts Discovery at five to ten per cent of spend. That figure assumes an account with four campaign types, where Discovery is one of several things competing for the money.
A new app running two campaigns is a different situation. Held at a tenth, a two-campaign account learns almost nothing new, because the only thing in it whose job is to find unknown terms is starved. Eighty / twenty is a more honest split for a starting account: four fifths to the terms you believe in, a fifth to finding out what you are missing.
As the account matures and the exact-match list fills with terms that have proved themselves, Discovery's share can come down — there is less left to discover.
It is not a cheaper way to run the same keywords, and it is not a set-and-forget campaign. A Discovery campaign that nobody harvests from is simply paying Apple to learn things nobody acts on. The value is not in running it; it is in reading its search terms and moving the winners across.
Yes. Search Match finds terms you would not have thought of, and Exact match is where those terms go once they prove they convert, at a bid you control. Running either alone gives up most of the value.
They will, unless you stop them. Every keyword your exact-match campaign carries should be added as a negative keyword in the Search Match campaign, or both bid on the same search and you pay more for traffic you already had.
Published guidance says five to ten per cent, but that assumes an account with four campaign types. For a new app running just two, a fifth is more sensible: at a tenth, the only campaign whose job is to find new terms is too starved to find any.
Broad match works from keywords you chose, extending them to plurals, misspellings and related phrases. Search Match uses no keyword list at all: Apple matches your app to searches using your App Store listing. Broad is an extension of your own list; Search Match is a search for what is missing from it.